Eclipse VC has launched a $1.3B fund dedicated to backing and incubating “physical AI” startups. With the humanoid robotics market projected to grow at a 45% CAGR, capital is shifting from pure software to embodied AI that interacts with the real world. For founders, this incubation model offers a critical pathway to overcome the high capital expenditure barriers traditionally associated with hardware startups.
The Shift from Cloud to Concrete: Why Physical AI is Winning
The launch of Eclipse VC’s $1.3B fund marks a definitive pivot in the venture landscape: the transition from pure software applications to “physical AI.” Physical AI encompasses AI-integrated robotics, embodied agents, and hardware systems that interact directly with the physical world. Driven by severe global labor shortages and the demand for industrial automation, the global AI hardware and robotics markets are projected to exceed $100 billion by 2030. More specifically, humanoid robotics is experiencing an explosive 45% compound annual growth rate (CAGR) from 2024 to 2030. In 2023 and 2024, robotics and physical AI subsets captured 15-20% of AI deal flow—up from just 5% in 2022. Investors are increasingly betting on real-world deployment over conversational chatbots, recognizing that the next trillion-dollar companies will build systems that move and act, not just text and chat.
The Incubation Advantage for Hardware Heavyweights
For founders, the most compelling aspect of Eclipse’s strategy is its focus on incubation. Historically, hardware startups have been notoriously difficult to fund due to massive early-stage capital expenditures (capex) required for prototyping, supply chain setup, and manufacturing. Eclipse’s model addresses this by building startups from the ground up, providing seed capital, engineering talent, and infrastructure before a company even needs to raise an external Series A. This model allows founders to avoid punishing early-stage dilution while tackling complex hardware-software integration challenges. It targets the massive gap left by incumbents like Boston Dynamics, bringing agile, venture-backed velocity to heavy engineering problems.
Competitive Landscape: Big Tech vs. Agile Startups
The physical AI arena is already highly competitive, dominated by massive funding rounds and strategic Big Tech partnerships. For instance, 1X Technologies recently raised a $100M Series B backed by OpenAI and Samsung to deploy enterprise and consumer humanoid robots. Adept, focusing on general intelligence for software-physical interfaces, achieved a $1B unicorn valuation after a $350M Series B backed by Microsoft and Nvidia. Anthropic continues to raise billions ($6.9B total) to build frontier models adaptable to physical systems. Meanwhile, startups like 01.AI are proving that efficient, edge-deployable large models are highly valued, reaching a $1B valuation in their first round. To compete, new founders must find distinct niches—such as Aionics, which focuses on AI for battery materials, enabling the very hardware that physical AI relies upon.
Strategic Playbook for Hardware-Software Founders
The influx of capital into physical AI presents both a massive opportunity and a high barrier to entry. Founders looking to capitalize on this wave must adopt specific strategies:
1. Bootstrap to a Physical Demo: Investors in this space need to see tangible interaction. Build Minimum Viable Products (MVPs) that demonstrate actual physical deployment and task execution, even if the hardware is off-the-shelf. The software-hardware integration is the moat.
2. Optimize for Edge Efficiency: As demonstrated by 01.AI, the ability to run complex models on edge devices (like a robot’s onboard computer) rather than relying on high-latency cloud connections is critical. Pitching cost controls and compute efficiency will extend your runway and attract investors.
3. Leverage Cross-Border Supply Chains: Capitalize on US venture funding while utilizing global supply chains. Partnering with strategic manufacturers in regions like South Korea (e.g., Samsung’s ecosystem) can reduce hardware costs by 20-30% and accelerate time-to-market.